How to Size Battery Banks for Off-Grid Systems

Undersized battery banks in off-grid systems lead to premature system failure during extended autonomy periods, with field data showing that ignoring depth of discharge can reduce usable capacity by 50% for AGM batteries. When engineers skip proper sizing calculations, installations experience voltage drop below inverter cutoff thresholds within hours rather than days, causing critical loads to fail and requiring emergency generator intervention. Oversizing materially increases capital cost (battery cost typically dominates an off-grid system budget) and adds weight and footprint that may violate NFPA 855 Section 4.5 installation clearance requirements for stationary energy storage.

Proper battery bank sizing prevents these failures by calculating the actual energy storage needed after accounting for real-world efficiency losses and discharge limitations. The fixed off-grid sizing model used in this calculator addresses the core engineering challenge: determining how much usable energy a system can deliver, not just how many batteries exist on paper. This approach aligns with Victron Energy's published guidance on practical depth of discharge differences between battery chemistries and IEEE 485 recommendations for lead-acid battery sizing in stationary applications.

Why Off-Grid Battery Sizing Goes Beyond Nameplate Ah

Battery bank sizing for off-grid systems is the process of determining the required nominal energy storage capacity to support electrical loads for a target autonomy period after accounting for system efficiency losses and usable depth of discharge limitations. This calculation converts load power requirements into actual battery specifications through energy balance equations that consider the complete power path from storage to load. The engineering necessity stems from the fundamental difference between nominal battery capacity and usable capacity in practical operation, where factors like inverter efficiency, wiring losses, and discharge limits reduce available energy.

Engineers need this calculation to comply with NFPA 855 Section 4.3.2, which requires energy storage systems to be sized based on actual load requirements and installation conditions. The calculation provides the quantitative basis for selecting battery chemistry, determining parallel/series configurations, and verifying that recharge sources can refill the bank within available time windows. Without this sizing, systems either fail to meet autonomy requirements or waste resources on excessive capacity.

The sizing process connects directly to system voltage selection: 48V systems require half the current of 24V systems for the same power, reducing cable losses by a factor of four (I²R scales with current squared) and improving overall efficiency in larger installations. This is why off-grid installations above ~2,000 W typically standardize at 48V rather than 24V.

The Energy-Balance Formula: From Load to Battery Count

E_load = P × t
E_adjusted = E_load / η
E_bank = E_adjusted / DoD
Ah_required = E_bank / V_system
Battery Count = ⌈Ah_required / Ah_battery⌉

P represents the instantaneous power demand at the load terminals in watts (W), with typical values ranging from 100W for small cabin lighting to 10,000W for telecom shelter equipment. t represents autonomy time in hours (h), ranging from 0.5 hours for short-term backup to 720 hours (30 days) for extended off-grid operation, defining the duration over which energy must be delivered without recharge.

η represents overall system efficiency as a decimal (typically 0.85 to 0.95), accounting for losses in inverters, charge controllers, and wiring between the battery bank and loads. η appears in the denominator because lower efficiency means more battery capacity is required to deliver the same load energy after losses. DoD represents usable depth of discharge as a decimal (typically 0.5 for AGM to 0.8 for LiFePO4), representing how much of nominal capacity can be safely cycled without accelerating cell damage or reducing cycle life.

V_system represents system voltage in volts (V), typically 12V, 24V, or 48V for off-grid installations, which converts energy requirements to amp-hours based on the fundamental relationship P = V × I. Ah_battery represents individual battery unit capacity in amp-hours (Ah), ranging from 50Ah for small deep-cycle batteries to 1000Ah for industrial cells, determining how many parallel units are needed to achieve the required capacity. The ceiling function in the battery count calculation ensures whole batteries are specified, preventing fractional units that don't exist in practical installations.

Remote Cabin: 250W Average Load on 24V AGM with 8h Autonomy

Consider a 400 square foot off-grid cabin with LED lighting (100W), refrigerator (150W), and occasional power tools (500W peak). The system must support 8 hours of autonomy at average load, using 24V architecture with AGM batteries limited to 50% depth of discharge and 85% system efficiency. Load power averages 250W continuous with occasional peaks handled by inverter surge capacity. Metric calculation: E_load = 250W × 8h = 2000Wh. E_adjusted = 2000Wh / 0.85 = 2353Wh. E_bank = 2353Wh / 0.5 = 4706Wh. Ah_required = 4706Wh / 24V = 196Ah. With 100Ah batteries: Battery Count = ⌈196Ah / 100Ah⌉ = 2 batteries in parallel.

Practical takeaway: at 196 Ah required, two 100 Ah AGM batteries in parallel give 200 Ah nominal — sufficient with no margin. Add at least one more 100 Ah unit (300 Ah nominal, 150 Ah usable at 50% DoD) for end-of-life capacity loss and temperature derating. Verify that the solar array delivers ~20 A at 24V on the design-day to refill the bank; otherwise the system runs in chronic partial state of charge, accelerating sulfation and shortening AGM lifespan to under 2 years per IEEE 1188 partial-SoC operation guidance.

Telecom Shelter: 5,000W on 48V LiFePO4 with 48h Hurricane Autonomy

A cellular tower shelter requires 5000W continuous power for radio equipment, cooling, and monitoring systems, with 48-hour autonomy mandated for hurricane-prone areas. The system uses 48V architecture with LiFePO4 batteries at 80% depth of discharge and 92% efficiency from high-quality components. Metric calculation: E_load = 5000W × 48h = 240,000Wh. E_adjusted = 240,000Wh / 0.92 = 260,870Wh. E_bank = 260,870Wh / 0.8 = 326,087Wh. Ah_required = 326,087Wh / 48V = 6793Ah. With 500Ah industrial cells: Battery Count = ⌈6,793 Ah / 500 Ah⌉ = 14 modules. Each 500 Ah module is built from 16 LiFePO4 cells in series at 3.2 V nominal (51.2 V actual, treated as 48 V system). The 14 modules connect in parallel through current-limiting fuses and an active battery management system per IEEE 1188 stationary battery sizing practices and NFPA 855 Section 4.5 arrangement requirements.

Practical takeaway: 326 kWh bank energy / 6,793 Ah at 48V is substantial storage that requires NFPA 855 Section 8.4 ventilation provisions even for LiFePO4 (off-gassing during overcharge faults remains a concern) and dedicated room layout to meet Section 4.5 clearance and access requirements. Comparison: an equivalent AGM bank at 50% DoD would need 1/0.5 / (1/0.8) = 1.6× more nominal capacity (~10,800 Ah at 48V) — much larger footprint, weight, and venting load. The recharge plan must deliver at least the average daily energy in the available solar window: 5,000 W × 24 h = 120 kWh per day at peak, requiring ~2,500 W of net solar charging power across the typical 5–6 hour design-day insolation period, plus generator backup sized for monsoon/hurricane recovery. NFPA 855 Section 7 requires both functions during commissioning.

What Drives Battery Bank Capacity Beyond the Formula

Depth of Discharge Limitations

Depth of discharge represents the most significant factor in battery bank sizing, with practical values differing substantially by chemistry. AGM batteries are typically held to 50% usable DoD in real cycling applications to prevent accelerated sulfation and extend cycle life — published cycle counts at 50% DoD typically reach 500–800 cycles per manufacturer datasheets. LiFePO4 routinely operates at 80% DoD with minimal degradation, with 2,000–5,000 cycle life depending on cell quality and operating temperature per Victron Energy and major Li-ion manufacturer published cycle data. A system requiring 10kWh usable storage needs 20kWh nominal capacity with AGM (10kWh / 0.5) but only 12.5kWh with LiFePO4 (10kWh / 0.8), representing a 37.5% reduction in physical batteries, weight, and footprint.

Temperature effects further modify effective depth of discharge, with battery capacity decreasing approximately 1% per °C below 25°C for lead-acid chemistries. At 0°C, an AGM battery's usable capacity may drop to 40% of nominal rather than 50%, requiring additional derating in cold climates. IEEE 485 provides correction factors for temperature effects in stationary applications, while NFPA 855 requires battery rooms maintain temperatures within manufacturer specifications. These interactions demonstrate why depth of discharge cannot be treated as a fixed percentage without considering installation environment and battery technology.

System Efficiency Losses

Overall system efficiency ranges from 85% to 95% in well-designed off-grid installations, with losses distributed across multiple components. Inverters typically operate at 90-95% efficiency at optimal load, while charge controllers add 2-5% losses depending on PWM vs MPPT technology. Wiring losses vary with cable sizing and distance, often adding 1-3% voltage drop that reduces effective energy delivery. A system with 85% efficiency requires 17.6% more battery capacity than one with 95% efficiency for the same load energy, making component selection critical for cost-effective sizing.

Efficiency varies with load percentage, with most inverters showing reduced efficiency below 20% of rated capacity. This nonlinear relationship means systems operating at light loads may experience efficiency drops to 80% or lower, requiring additional capacity margin. The calculator's fixed efficiency assumption provides a conservative estimate, but engineers should verify actual efficiency curves for specific equipment, especially when loads vary significantly. Proper cable sizing per NEC Article 310 reduces wiring losses, while selecting MPPT charge controllers over PWM types improves solar charging energy harvest by 15–30% per manufacturer comparative data — the gain is largest when array Vmp differs significantly from battery voltage (cold-temperature spike or partial shading).

System Voltage Selection

System voltage determines the current required for a given power level, directly affecting cable sizing, losses, and component costs. A 5000W load requires approximately 417A at 12V (5000W / 12V), 208A at 24V, and 104A at 48V, with cable cross-sectional area increasing proportionally to current squared for the same voltage drop. Higher voltage systems reduce I²R losses significantly: a 48V system experiences one-fourth the wiring losses of a 24V system for the same power and cable resistance. This efficiency gain reduces required battery capacity by minimizing energy wasted in distribution.

Voltage selection also affects battery configuration, with 48V systems typically requiring four 12V batteries in series or individual LiFePO4 cells arranged in 16S configurations. Series connections introduce balancing requirements and potential single-point failures, while parallel connections require careful current sharing to prevent uneven aging. NFPA 855 Section 4.5 addresses battery arrangement and protection requirements for different configurations. The calculator's system voltage input allows engineers to compare alternatives, with higher voltages generally preferred for systems above 2000W due to reduced losses and smaller conductor requirements.

Where the Energy-Balance Formula Falls Short

The E_load → E_adjusted → E_bank chain is a steady-state energy balance. Five conditions push real off-grid sizing beyond what the formula captures:

1. Surge and inrush loads. The formula uses average load × time. It does not size for inrush currents from compressors, well pumps, or induction motors that demand 4–8× rated power for 0.1–1 second on each start. Inverter surge rating (typically 2–3× continuous rating for 5 seconds) handles short transients, but the battery must be able to supply the surge current without voltage collapse. For motor-heavy loads, separately verify battery instantaneous discharge rate (C-rate) and inverter surge capacity per IEEE 1188.

2. End-of-life capacity loss. Manufacturer Ah ratings reflect new cells. AGM reaches 80% capacity at end of useful life (5–7 years); LiFePO4 reaches 80% at 2,000–5,000 cycles depending on temperature and DoD. Add 1.25× margin for AGM and 1.10× margin for LiFePO4 to maintain autonomy across the design lifespan.

3. Temperature derating. IEEE 485 Annex C provides correction factors for cold operation: at -10°C, AGM usable capacity drops to ~70% of 25°C nominal. LiFePO4 charging is restricted below 0°C without integrated cell heaters; many BMS units inhibit charge entirely below freezing. Cold-climate installations need either heated battery enclosures or further capacity oversizing.

4. Partial state-of-charge degradation. Lead-acid chemistries suffer accelerated sulfation if held below 80% SoC for extended periods. Off-grid systems often operate in partial SoC during winter or extended cloud cover. The formula sizes for autonomy but does not check whether typical recharge cycles bring the bank back above 80% — a separate recharge calculation against PV array output and generator runtime is required, as covered in the recharge verification step.

5. Round-trip efficiency split. The η in the formula typically represents the discharge path (battery → inverter → AC load). Charge path efficiency (PV array → MPPT → battery) is separate: ~95% for LiFePO4, ~80–85% for AGM (charge acceptance limits absorb less of input energy). Total system round-trip efficiency for AGM-based off-grid is typically 70–75%, for LiFePO4 85–90%. When sizing PV array and generator runtime, use the chemistry-specific charge efficiency, not the discharge η used in the bank sizing formula.

Where Battery Bank Sizing Goes Wrong

Engineers frequently size batteries based solely on amp-hour ratings without first calculating required energy storage over the autonomy period. This mistake occurs when selecting batteries before determining load profiles, leading to systems that may provide sufficient current but insufficient runtime. For example, specifying a 200Ah battery bank at 12V (2400Wh nominal) for a 500W load assumes 4.8 hours runtime at 100% efficiency and DoD, but actual usable capacity may be only 1200Wh at 50% DoD, providing just 2.4 hours autonomy. The field consequence is premature system shutdown during extended outages, requiring expensive retrofits or generator dependency that violates off-grid design intent.

Applying identical depth of discharge across battery chemistries is another common error. There are two patterns: (1) Holding 50% DoD when substituting LiFePO4 for AGM oversizes the bank by 60% — a 100 Ah AGM at 50% DoD = 50 Ah usable, but 100 Ah LiFePO4 at the same 50% DoD assumption gives 50 Ah usable when the chemistry actually supports 80 Ah usable safely. The engineer ends up with too many cells, capital cost overrun, and unnecessary footprint. (2) Assuming LiFePO4 supports 100% DoD undersizes the bank by 20% — practical LiFePO4 at 80% DoD per Victron Energy guidance and IEEE 1188 still requires headroom for BMS protection thresholds, temperature derating, and cell aging. Use chemistry-specific DoD values: 50% for AGM/flooded lead-acid, 80% for LiFePO4, 70% for nickel-iron — and never carry over a DoD value when changing chemistry.

Neglecting to verify recharge capability after sizing the battery bank causes systems to become partially charged between discharge cycles, leading to progressive capacity loss. A bank sized for 48-hour autonomy may require 100A charging current to recharge within 8 sunlight hours, but the solar array or generator may only provide 50A, creating a daily deficit that accumulates over time. This mistake violates IEEE 1188 recommendations for maintaining batteries above 80% state of charge between cycles and leads to premature failure within months rather than years. Engineers must perform separate recharge calculations using available charging sources, accounting for seasonal variations and weather patterns that affect solar harvest or generator runtime.

Sizing Multipliers and Recharge Verification Workflow

For reliable off-grid operation, nominal battery bank capacity must exceed the bare load energy by a minimum factor of about 2.35 for AGM chemistry or 1.47 for LiFePO4 based on the energy-balance formula alone (1/0.85 × 1/0.5 for AGM, 1/0.85 × 1/0.8 for LiFePO4). Add an additional 6–10% margin for end-of-life capacity loss (per IEEE 485 stationary battery aging derating) and 5–10% for temperature derating in cold installations (per manufacturer temperature correction curves), bringing practical sizing factors to roughly 2.6 for AGM and 1.6–1.7 for LiFePO4. Systems that fall below the formula-only factor risk autonomy failure on day one; systems that hit only the formula-only factor without aging margin reach autonomy failure within 3–5 years as batteries lose nominal capacity.

Use the battery bank sizing calculator during preliminary design to establish baseline requirements, then refine with specific equipment efficiencies and manufacturer depth of discharge data before final specification. The calculated Ah_required value guides battery selection and configuration, while the battery count determines physical layout and balance of system components. Validate results against recharge source capacity using separate calculations, ensuring the bank can return to full charge within available time windows between discharge events. This complete sizing approach ensures off-grid systems meet autonomy requirements while optimizing cost, weight, and space utilization across the system lifecycle.

FAQ

Related Calculators

• Battery Bank Sizing for Off-Grid
• Solar Charge Controller Sizing Calculator
• Cable Ampacity Calculator
• Voltage Drop Calculator
• UPS Battery Runtime Calculator